Granica, which helps AI companies optimize their cloud object storage in Amazon S3 and Google Cloud, emerges from stealth with $45M from NEA, BCV, and others
Belle Lin / Wall Street Journal :
Context & Ripple Effects
Granica is entering a niche with a proven exit template: Granulate, which applied the same optimize-cloud-infrastructure-with-AI playbook to compute, went from a $30M Series B to a $650M acquisition by Intel on roughly $45M of total funding. NEA and BCV are now backing the storage-layer version of that thesis as AI companies' S3 and Google Cloud bills balloon.
The bet lands amid hyperscalers pouring tens of billions into AI capacity — Amazon, Microsoft, and Google alone pledged a combined $67.5 billion for India since October — which makes every percentage point of storage efficiency a sellable product sitting on top of their own infrastructure.
First-order effects
- AI companies running training and inference data on Amazon S3 and Google Cloud gain a dedicated cost-optimization vendor, directly attacking a line item those same customers are scaling fastest.
- NEA and BCV get early positioning in storage optimization, a category whose closest analog, Granulate, returned more than ten times its raised capital to investors when Intel bought it.
Second-order effects
- Amazon and Google must decide whether to treat Granica as a partner shaving waste off their storage revenue or a threat monetizing their pricing opacity — the Granulate-Intel outcome suggests chipmakers and platform owners may simply acquire rather than compete.
- Rival cloud-cost tooling vendors face pressure to add AI-workload-specific optimization or cede the fastest-growing segment of the bill to a purpose-built entrant.
Third-order effects
- If the Granulate-to-Intel arc repeats, expect a standing M&A lane where efficiency layers built on hyperscaler infrastructure become acquisition targets for the platforms themselves, structuring venture returns around exits to AWS, Google, and chip vendors.
- For AI labs, storage shifts from a raw utility toward a negotiated, instrumented cost center, with third-party optimizers inserting themselves between buyers and hyperscaler list prices.
The trend: The AI capex surge is spawning a venture-funded layer of startups that arbitrage hyperscaler infrastructure costs, with platform and chipmaker acquisitions as the likely endgame.